A Performance Analysis of Different Classification Techniques in Offline Handwritten Signature Verification

Offline handwritten signature system works on the scanned image of a signature. Offline handwritten signature verification is a two class pattern recognition problem. For our experimentation purpose, we have developed offline signature datasets of with genuine and forged signature samples. Some commonly used geometric features were extracted from the signature datasets. Sequential Minimization Optimization algorithm with different kernels and Naive Bayes were used as the classification techniques. Performance analysis of different classification techniques is also discussed in terms of False Acceptance Rate (FAR) and False Rejection Rate (FRR).